Biocompatibility and intradiscal application of a ...Introduction Chronic low back pain is a debilitating disorder associ-ated with intervertebral disc (IVD) degeneration [1]. As the

Willems et al. Arthritis Research & Therapy (2015) 17:214 DOI 10.1186/s13075-015-0727-x

RESEARCH ARTICLE Open Access

Biocompatibility and intradiscal applicationof a thermoreversible celecoxib-loadedpoly-N-isopropylacrylamide MgFe-layereddouble hydroxide hydrogel in a canine modelNicole Willems1, Hsiao-yin Yang2, Marloes L. P. Langelaan3, Anna R. Tellegen1, Guy C. M. Grinwis4,Hendrik-Jan C. Kranenburg1, Frank M. Riemers1, Saskia G. M. Plomp2, Eric G. M. Craenmehr3, Wouter J. A. Dhert1,2,Nicole E. Papen-Botterhuis3, Björn P. Meij1, Laura B. Creemers2† and Marianna A. Tryfonidou1*†

Abstract

Introduction: Chronic low back pain due to intervertebral disc (IVD) degeneration is associated with increasedlevels of inflammatory mediators. Current medical treatment consists of oral anti-inflammatory drugs to alleviatepain. In this study, the efficacy and safety of a novel thermoreversible poly-N-isopropylacrylamide MgFe-layereddouble hydroxide (pNIPAAM MgFe-LDH) hydrogel was evaluated for intradiscal controlled delivery of the selectivecyclooxygenase (COX) 2 inhibitor and anti-inflammatory drug celecoxib (CXB).

Methods: Degradation, release behavior, and the ability of a CXB-loaded pNIPAAM MgFe-LDH hydrogel to suppressprostaglandin E2 (PGE2) levels in a controlled manner in the presence of a proinflammatory stimulus (TNF-α) wereevaluated in vitro. Biocompatibility was evaluated histologically after subcutaneous injection in mice. Safety ofintradiscal application of the loaded and unloaded hydrogels was studied in a canine model of spontaneous mildIVD degeneration by histological, biomolecular, and biochemical evaluation. After the hydrogel was shown to bebiocompatible and safe, an in vivo dose–response study was performed in order to determine safety and efficacy ofthe pNIPAAM MgFe-LDH hydrogel for intradiscal controlled delivery of CXB.

Results: CXB release correlated to hydrogel degradation in vitro. Furthermore, controlled release from CXB-loadedhydrogels was demonstrated to suppress PGE2 levels in the presence of TNF-α. The hydrogel was shown to exhibita good biocompatibility upon subcutaneous injection in mice. Upon intradiscal injection in a canine model, thehydrogel exhibited excellent biocompatibility based on histological evaluation of the treated IVDs. Gene expressionand biochemical analyses supported the finding that no substantial negative effects of the hydrogel were observed.Safety of application was further confirmed by the absence of clinical symptoms, IVD herniation or progression ofdegeneration. Controlled release of CXB resulted in a nonsignificant maximal inhibition (approximately 35 %) ofPGE2 levels in the mildly degenerated canine IVDs.

Conclusions: In conclusion, this study showed biocompatibility and safe intradiscal application of an MgFeLDH-pNIPAAM hydrogel. Controlled release of CXB resulted in only limited inhibition of PGE2 in this modelwith mild IVD degeneration, and further studies should concentrate on application of controlled release fromthis type of hydrogel in animal models with more severe IVD degeneration.

* Correspondence: [email protected]†Equal contributors1Department of Clinical Sciences of Companion Animals, Faculty ofVeterinary Medicine, Utrecht University, Yalelaan 108, Utrecht 3584CM, TheNetherlandsFull list of author information is available at the end of the article

© 2015 Willems et al. Open Access This artiInternational License (http://creativecommoreproduction in any medium, provided youlink to the Creative Commons license, andDedication waiver (http://creativecommonsarticle, unless otherwise stated.

cle is distributed under the terms of the Creative Commons Attribution 4.0ns.org/licenses/by/4.0/), which permits unrestricted use, distribution, andgive appropriate credit to the original author(s) and the source, provide aindicate if changes were made. The Creative Commons Public Domain.org/publicdomain/zero/1.0/) applies to the data made available in this

http://crossmark.crossref.org/dialog/?doi=10.1186/s13075-015-0727-x&domain=pdfmailto:[email protected]://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/

Willems et al. Arthritis Research & Therapy (2015) 17:214 Page 2 of 16

IntroductionChronic low back pain is a debilitating disorder associ-ated with intervertebral disc (IVD) degeneration [1]. Asthe exact pathogenesis is still poorly understood, currentsurgical and medical treatments aim at alleviating symp-toms. Inhibiting or reversing the degenerative process byusing advanced methods like cell and tissue engineeringare in development, but are not clinically applicable thusfar. In degenerative disc diseases, the specific compos-ition of the nucleus pulposus (NP) and annulus fibrosus(AF) is disturbed, since the delicate equilibrium shifts to-ward the catabolic pathways [2, 3]. In the NP this resultsin a change from an extracellular matrix (ECM) rich inproteoglycans and type II collagen, to a tissue containingmainly type I collagen, and in the AF in a loss of lamel-lar organization [4, 5]. The loss of proteoglycans causesa decrease in the water-binding capacity of the NP andtogether with the changes in the AF, compromises thestructural functionality of the IVD [5].A variety of inflammatory mediators has been investi-

gated for their role in the catabolic processes of IVD de-generation; targeting the inflammation process is one ofthe emerging treatment strategies of chronic low backpain and IVD degeneration. Herniated degenerative disctissue has been shown to spontaneously produce in-creased amounts of matrix metalloproteinases (MMPs),nitric oxide, prostaglandin E2 (PGE2) and interleukin 6(IL-6), and to express interleukin 1 (IL-1), interleukin 8(IL-8), tumor necrosis factor alpha (TNF-α) [6, 7]. TNF-αand IL-1 upregulate expression of matrix-degrading en-zymes by NP cells [4, 8, 9]. Furthermore, elevated levels ofIL-1 and PGE2 have been associated with aging and de-generation of the IVD [9, 10]. In the NP, PGE2 negativelyaffects matrix integrity by inhibiting proteoglycan synthe-sis, possibly mediated by a decrease in insulin growth fac-tor 1 and an increase in matrix-degrading enzymes [10].PGE2 is a well-known prostanoid and plays an import-

ant regulatory role in physiological as well as pathologicalprocesses like intervertebral disc degeneration. It is syn-thesized by two cyclooxygenase (COX) isoforms, COX-1and COX-2, by conversion of arachidonic acid into prosta-glandin H2 (PGH2) and isomerization of PGH2 to PGE2by prostaglandin E synthases (PTGES). COX-1 is constitu-tively expressed in most tissues and is associated with theproduction of baseline PGE2 levels important for homeo-stasis. In contrast, COX-2 expression is highly restrictedunder physiological conditions, but can be rapidly inducedin response to inflammatory stimuli and is therefore be-lieved to play an important role in the PGE2 productioninvolved in degenerative processes [11, 12]. SelectiveCOX-2 inhibitors have been developed to reduce PGE2production via this pathway. In several clinical trials the ef-ficacy of COX-2 inhibitors in patients with low back painhas been established [13, 14]. However, their widespread

application is hampered by severe side effects, such ascardiotoxicity. Although these inhibitors can be effectivelyintroduced into the avascular IVD by intradiscal injection,they would achieve only short-lived clinical effects. De-livering drugs by using controlled release systems, e.g.,hydrogels, would be a more attractive alternative to bolusinjections, as a higher loading dose and long-term deliverycan be accomplished by a minimum of intradiscal inter-ventions [15, 16].Temperature-sensitive poly-N-isopropylacrylamide

(pNIPAAM)-based hydrogels have been extensively usedin the field of controlled release [17]. These gels could beparticularly suitable for intradiscal injection as they are li-quid at room temperature, and hence injectable throughsmall-diameter needles and form a solid gel at 37 °C, pre-venting leakage of injected materials from the IVD [18]. Inthis study a hybrid thermoreversible biodegradable hydro-gel served as a controlled release platform for the spe-cific COX-2 inhibitor celecoxib (CXB). This releasesystem consists of lower critical solution temperature(LCST) polymers with a low molecular weight, based onpNIPAAM with a sulfonate end group, ionically linked to anetwork of biodegradable platelet-type MgFe layereddouble hydroxide (LDH) nanoparticles. LDH particlespossess a positive surface charge, which can interactwith the negatively charged LCST polymers. At roomtemperature these hybrid structures are simply made bymixing the polymer with the MgFe-LDHs and subsequentdispersion in water. This results in a solution of low vis-cosity that can easily be injected into the NP via a 29Gneedle. At body temperature (37 °C), physical entangle-ments are formed due to hydrophobic interactions be-tween the LCST polymers, resulting in the formation of ahybrid network, as polymers are linked to the MgFe-LDHparticles. Furthermore, the easily ionizable carboxylicgroups of CXB can interact with the biodegradable LDHs,which makes this unique for drug release. We hypothesizethat pNIPAAM MgFe-LDH hydrogels are suitable vehiclesfor delivering a COX-2 inhibitor into the IVD, to reduceintradiscal PGE2 levels over time in a dog model withspontaneous IVD degeneration, showing pathophysio-logical aspects similar to those in human [19, 20].

Materials and methodsAfter synthesis, preparation, and rheological analysis ofthe hydrogel, the degradation and release behavior of theCXB-loaded hydrogel was evaluated. Furthermore, in anin vitro model in the presence of a proinflammatory stimu-lus (TNF-α) the ability of the CXB-loaded hydrogels tosuppress PGE2 levels in a sustained manner was evaluated.Thereafter, biocompatibility upon subcutaneous injectionwas studied in mice. Safety of intradiscal application of theloaded and unloaded hydrogel was studied in a caninemodel of spontaneous mild IVD degeneration. After the

Fig. 1 Rheological setup and results of the complex shear modulus|G*| of the pNIPAAM MgFe-LDH hydrogel. a Configuration of therheological setup to measure |G*| of the pNIPAAM MgFe-LDHhydrogel. b |G*| of the pNIPAAM MgFe-LDH hydrogel as a functionof temperature. The dashed red line connects the time point wherethe |G*| starts to increase to the temperature curve, and indicates alower critical solution temperature (LCST) of 32 °C. AS after sterilization,ASC after sterilization separate components, NS nonsterilized. G*complex shear modulus, LDH layered double hydroxide, pNIPAAMpoly-N-isopropylacrylamide


hydrogel was shown to be biocompatible and safe, afollow-up dose–response in vivo study was performed inorder to determine safety and efficacy of the pNIPAAMMgFe-LDH hydrogel for intradiscal controlled delivery ofCXB.

Synthesis and preparation of pNIPAAM MgFe-LDHhydrogelsThe poly-N-isopropylacrylamide (pNIPAAM) polymer withsulfonate end group was synthesized as reported previously[21] and the modified synthesis is described in detail inAdditional file 1. To formulate the hydrogel, the pNIPAAMpolymer was added to the LDH suspension in a vial andsubsequently placed on a tube roller mixer for 48 h at roomtemperature and sterilized by gamma radiation (25 kGy,Isotron Nederland BV, Ede, The Netherlands). The finalhydrogel contained 16 wt % pNIPAAM, 3.3 wt % MgFeLDH and water.

Rheological analysis of the pNIPAAM MgFe-LDHhydrogelsThe viscoelastic properties of the unloaded pNIPAAmMgFe-LDHs were determined by using an Anton PaarMCR301 rheometer (Anton Paar Ltd., St. Albans, UK) withan oscillatory parallel plate geometry (50 mm diameter)with a constant strain of γ = 0.5 % at a frequency of f = 1Hz. Temperature was increased from 22 °C to 37 °C at arate of 15 °C/min. This heating rate was chosen based oncalculating the minimum rate of heat transfer based onestimating the energy needed to heat up the hydrogel,by taking into account the surface of the hydrogel andthe minimum temperature difference between the LCSTand body temperature. LCST is the critical temperatureabove which the hydrogel undergoes a phase transitionfrom a soluble to an insoluble state. This estimation isdescribed in detail in Additional file 1. The gelling wasrecorded by measuring the complex shear modulus |G*|,which is a common parameter to determine the strengthof a viscoelastic material like a hydrogel. The complexshear modulus |G*| is correlated to the storage modulus(G’) and loss modulus (G”). The storage modulus is ameasure of the deformation energy stored in the sampleduring the shear process (elastic behavior), whereas theloss modulus is a measure of the energy dissipated in thesample during the shear process (viscous behavior), and islost to the sample afterward (viscous behavior). The rela-tion between these parameters is the following:

G�j j ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiG0ð Þ2 þ G}ð Þ2

q

Hydrogel samples were placed on the lower plate, andthe upper plate was lowered to a 0.5 mm gap. The con-figuration of the rheological setup is shown in Fig. 1a.The viscoelastic properties of the loaded hydrogel were

not determined. The CXB concentrations were as low as10−6 M to 10−4 M, i.e., 0.38–38 mg of celecoxib per liter,or 0.0038–3.8 × 10-5 wt %, and were not influencing therheological properties.

Degradation and release behavior of CXB-loaded pNIPAAMMgFe-LDH hydrogelsIn vitro, the controlled release of CXB from hydrogelswas measured in phosphate-buffered saline (PBS) (pH 7.4,44 mM Na2HPO4, 9 mM NaH2PO4, 72 mM NaCl, 0.02 %wt NaN3) and 0.2 % Tween 80® (polyoxyethylenesorbitanmonooleate; Sigma-Aldrich Chemie B.V., Zwijndrecht,The Netherlands). Tween 80® was added to the buffer inorder to increase the solubility of CXB [22] and thereby simu-late the in vivo situation. CXB-loaded pNIPAAM MgFe-LDHsuspension was prepared by adding 6 or 10 mg/ml ofCXB to the dispersion and stirring with a stirring barfor 2 days. A volume of 1 ml of CXB-loaded pNIPAAMMgFe-LDH suspension was pipetted into a vial and placedfor 30 min at 37 °C to ensure gelation of the hydrogel, andafterward covered with 14 ml warm (37 °C) PBS/Tween80® solution. The release experiment was performed at37 °C. At day 1, 2, 5, 8, 15, 22, and 31, 12 ml of the buf-fer solution was removed in order to analyze CXB andMg concentrations and 12 ml of fresh buffer was added.


CXB concentrations were determined in a volume of100 μl by using ultra-performance liquid chromatography(UPLC) as described in detail recently by Petit et al. [23].In vitro degradation was determined by measuring therelative cumulative release of Mg into the medium. Mgconcentrations were determined by using a Prodigy HighDispersion Inductively Coupled Optical Emission Specto-metry (ICP-OES) system (Leeman Labs, St Charles, IL,USA). Standards were prepared by using multi-element(23 elements in diluted nitric acid) standard solution IV(1000 mg/l) (Merck Millipore, Darmstadt, Germany). Avolume of 0.5 ml of the solutions from the degradation ex-periment was diluted in 100 ml aqueous 1N HNO3, andsubsequently diluted again tenfold in 1N HNO3. The ef-fects of CXB loading (10 mg/ml versus 6 mg/ml), LDHcontent (single versus double) and type of LDH (Mg3Feversus Mg2.5Fe) of the gels on release behavior andin vitro degradation were also investigated, as well asCXB solubility effects by using a buffer containing PBSand 0.2 % or 2 % Tween 80®.

Controlled release of CXB in vitroThe balance between anabolic and catabolic pathways inarticular chondrocytes as well as NP cells can be directedtoward catabolism by TNF-α [24, 25]. Bovine articularchondrocytes were used in the in vitro experiments, asthey were more easily available in our laboratory. Articularchondrocytes were isolated from bovine carpometacarpaljoints by enzymatic digestion overnight with 2 mg/mlcollagenase A (Roche Diagnostics Deutschland GmbH,Mannheim, Germany) at 37 °C. Chondrocytes were seededat 5 × 106 cells/ml density (P0) into cylindrical (diameterand height 6 mm) 2 % agarose (Type VII, Sigma-AldrichChemie B.V.) constructs and left to gel at roomtemperature. The constructs were then cultured in 12-wellplates in high-glucose Dulbecco’s modified Eagle’s medium(hgDMEM; Gibco, Life Technologies Europe, Bleiswijk,The Netherlands) with 20 % fetal bovine serum (GreinerBio-One, Alphen aan Den Rijn, The Netherlands), 0.1 %amphotericin (Sigma-Aldrich Chemie B.V.), 1 % Pen-Strep(Biochrom GmbH, Berlin, Germany), 1 % nonessentialamino acids (Lonza, Basel, Switzerland) 1 % essentialamino acids (Lonza), and 50 mg/ml ascorbate 2-phosphate(Biochrom). During the first 5 days of culturing, these con-structs were stimulated with 10 ng/ml TNF-α to induce aninflammatory response reflected by elevated PGE2 levels(Fig. 2a). A concentration of 1 μM CXB has been describedto effectively lower PGE2 levels in osteoarthritic chondro-cytes and corresponds with mean pharmacological plasmalevels [26]. CXB was dispersed in the pNIPAAM MgFe-LDH mixture at a concentration of 0.1 mg/ml, aiming toestablish a concentration of approximately 1 μM CXB perculture medium renewal over the 28-day culture period.Controlled release of the CXB is achieved by dissolution of

the CXB crystals present a depot within the hydrogel anddiffusion of the solubilized CXB. A volume of 100 μl ofthe hydrogel suspension was pipetted at the bottom ofa 12-well plate and placed in an incubator at 37 °C toensure gelation of the hydrogels. Subsequently, definedas day 0 of the experiment, cell constructs and culturemedium were added to the 12-well plate. For the “bolusinjection” of CXB, only cell constructs were placed onthe bottom of the well, and CXB was added to the mediumevery 2 days, starting at day 0, at a concentration of 1 μM.Media were renewed every 2–3 days, collected on days 0,2, 7, 9, 11, 14, 21, and 28 and stored at −80 °C for analysisof CXB content. For the in vivo experiments based on for-mulation of higher doses of CXB, in vitro experimentswere also carried out with a higher dosage of CXB toevaluate release profiles at higher dosing. To this end, 1mg of CXB-loaded per ml of hydrogel, aiming to establisha concentration of 10 μM CXB per culture medium re-newal was used. Conditioned media were analyzed forCXB content and PGE2 levels. Inhibition of COX-2 ac-tivity was determined by measuring PGE2 in culturemedium. A colorimetric competitive enzyme immuno-assay kit (PGE2 EIA kit, Enzo Life Sciences BVBA,Antwerp, Belgium) was used to determine PGE2 levelsin culture medium according to the manufacturer’sinstructions.

In vivo biocompatibility in mice after subcutaneousimplantationAll animal procedures were approved and performed inaccordance with the guidelines set by the Animal Experi-ments Committee (DEC) of Utrecht University (experi-ment numbers: DEC 2010.III.03.046; DEC 2012.III.05.046;and DEC 2013.III.02.017). Six healthy female adult (8–10weeks old) BALB/c mice (Harlan-Olac Ltd., Bicester, UK)were used for testing biocompatibility and biosafety of thepNIPAAM MgFe-LDH polymer hydrogel and seven otherbiomaterials. Four different biomaterials were injectedat least 1 cm apart from each other into the dorsal sub-cutaneous tissue of each mouse in a randomized fashion.Buprenorphine 100 μg/kg was given intraperitoneally (i.p.)as premedication and analgesic and subsequently all ani-mals were anesthetized with isoflurane via an inductionmask (vaporizer setting 2.5 %) in a 1:1 oxygen:air mixture.A blood sample was drawn to perform a white blood cellcount and differentiation at day 0, to rule out systemic in-flammation. A volume of 200 μl of each biomaterial wasinjected subcutaneously with a 27G needle under sterileconditions. PBS (200 μl) served as a control. All injectionsites were marked with a waterproof marker. Immediatelyafter injection, Dermabond® (Ethicon, Cornelia, GA, USA)was applied to the injection site to prevent leakageand the injection site was heated by an infrared lampfor 1 min. Mice were monitored daily for signs of distress

Fig. 2 In vitro controlled release of CXB results into sustained suppression of PGE2 levels in the presence of a proinflammatory stimulus. a Experimentalsetup to evaluate the controlled release of celecoxib (CXB) in vitro. During the first 5 days of culturing, three-dimensional chondrocyte constructs werestimulated with a proinflammatory cytokine TNF-α (10 ng/ml). At day 0 of the experiment, a 10 μM bolus injection of CXB was applied for 2 consecutivedays or pNIPAAM MgFe-LDH hydrogels loaded with 1 mg/ml CXB (CR CXB). Media were refreshed every 2–3 days and collected on days 0, 2, 7, 9, 11, 14,21, and 28. On days 14 and 28 the constructs were evaluated for cell viability. b Celecoxib (CXB) concentrations (μM) were measured in medium samplesafter administering the bolus injection of CXB or the pNIPAAM MgFe-LDH hydrogel loaded with 1 mg/ml CXB (CR CXB) c. PGE2 concentrations (ng/ml)measured in medium samples in the following conditions: unstimulated, stimulated with 10 ng/ml TNF-α (TNF-α stimulated), TNF-α stimulatedin the presence of a 10 μM bolus injection of celecoxib (CXB bolus) for 2 consecutive days, or 1 mg/ml CXB-loaded pNIPAAM MgFe-LDHhydrogels (CR + CXB). Data are expressed as mean ± standard deviation. CR control, LDH layered double hydroxide, PGE2, prostaglandin E2,pNIPAAM poly-N-isopropylacrylamide, TNF-α tumor necrosis factor alpha


or pain (e.g., lethargy, weight loss, automutilation, andabnormal posture) and injection sites were monitoredfor inflammation (e.g., swelling, redness, pain, and heat).Three animals were sacrificed 7 days after injection, andthree after 28 days. At the end of the experimental period,animals were anesthetized with isoflurane, blood was col-lected by cardiac puncture for white blood cell count anddifferentiation, and euthanasia was performed by cervicaldislocation. The injection sites were removed for histo-logical analysis. Tissues were fixated in a 4 % neutral-buffered formaldehyde solution (Klinipath B.V., Duiven,The Netherlands) and after fixation routinely embeddedin paraffin. Sections of 4 μm were stained with hematoxylinand eosin. Infiltration of inflammatory cells, giant cells,necrosis, neovascularization, fatty infiltration, and theencapsulation of the biomaterial by a fibrotic capsulewere histologically assessed as parameters for a bio-logical response at the application site, at 7 and 28 daysby a blinded board-certified veterinary pathologist (GG)and the principal investigator (NW) using an OlympusBX41 microscope (Olympus Europa GmbH, Hamburg,Germany).

Intradiscal application of CXB-loaded pNIPAAM MgFe-LDHhydrogels in laboratory beagle dogsData from two in vivo studies in beagle dogs were com-bined and analyzed. Both studies were set up as random-ized block designs. In the first study CXB-loaded (7.7 μM)and unloaded pNIPAAM MgFe-LDH hydrogels, a bolusinjection of CXB (7.7 μM) and 0.9 % NaCl were intradis-cally injected. In other levels two other materials irrelevantto this study were injected. The second study served as adose–response study, including a 10- and 100-fold higherdosage of CXB (77 μM and 770 μM) in addition to the7.7 μM dose. For preparation of the CXB-loaded hydro-gels, CXB was prepared from a CXB stock solution inethanol (60 μg/ml) by sterile filtration. Water was addedto this ethanolic solution of CXB to obtain a dispersionwith small CXB crystals (diameter approximately 1 μm).This dispersion was freeze-dried overnight and thepNIPAAM MgFe-LDH mixture was added and incubatedovernight on a tube roller mixer at room temperature.In total 18 intact female beagle dogs (Harlan, Gannat,

France) with a median age of 1.7 years (range 1.3–1.8years) and a median weight of 8.4 kg (range 6.2–13.8 kg)


were used. Nine dogs with a median age of 1.6 years(range 1.3–1.8 years) and a median weight of 8.2 kg(range 6.2–11 kg) were used in the first study. Nine dogswith a median age of 1.7 years (range 1.6–1.8 years) anda median weight of 9.3 kg (range 8.3–13.8 kg) were usedin the second study. All dogs underwent general, ortho-pedic, and neurologic examination by a board-certifiedveterinary surgeon (BM).

Surgical procedureTo determine the grade of degeneration of the IVDsprior to surgery, magnetic resonance (MR) images of thelumbar vertebral column were obtained in fully anesthe-tized dogs. A blood sample was drawn from the jugularvein to assess white blood cell count and differentiation,to exclude systemic inflammation. Dogs were placed in adorsal recumbent position and throughout the completescan protocol heart rate, respiration rate, temperature, car-bon dioxide, and oxygen levels were monitored. The MRimaging was performed using a 0.2 Tesla open magnet(Magnetom Open Viva, Siemens AG, Munich, Germany).All lumbar IVDs were assessed according to the Pfirrmannscore by a veterinary radiologist on sagittal T2-weightedfast spin echo (FSE) images (3.0 mm slices, repetition time(TR) 4455 ms, echo time (TE) 117 ms) [27]. Only lumbarIVDs with a Pfirrmann score II were included for injection.The anesthesia protocol during surgery was similar

to the one used during MR scanning. Analgesia wasprovided by a combination of fentanyl (loading dose10 μg/kg, 15–20 μg/kg/h continuous rate infusion, c.r.i.)and ketamine (0.5 mg/kg loading dose, 10 μg/kg/minc.r.i.) intravenous (i.v.). Throughout the complete pro-cedure heart rate, respiration, temperature, carbon di-oxide, oxygen levels, and blood pressure (noninvasive)were monitored. Surgical sites were prepared accordingto standard protocol. A detailed description of the surgicalprocedure has been described previously [28]. Briefly, dogswere positioned in a right recumbent position to exposeand inject the T13-L1 until L6-L7 via a left lateral ap-proach. To diminish injury of the iliopsoas muscle and sci-atic nerve traction injury, the surgical approach in thesecond study was adjusted and L6-L7 and L7-S1 wereinjected via a dorsal approach, while the dogs were posi-tioned in ventral recumbency. In the first study a 100 μlsyringe (7638–01 Model 710 RN, Hamilton CompanyUSA, Reno, NV, USA), and in the second study a 100 μlgastight syringe (7656–01 Model 1710 RN) with a 29Gneedle (25 mm, 12° beveled point; Hamilton CompanyUSA) was used to inject 30 μl of the earlier mentionedcompounds through the AF into the NP. The smallestpossible needle diameter was chosen to minimize injuryto the treated IVDs. Wound closure was performed ac-cording to standard protocol. Postoperative pain man-agement in all dogs consisted of methadone 0.3 mg/kg

intramuscular (i.m.) quaque (q).6.h. during the first 24 hpostoperatively and buprenorphine 20 μg/kg i.m. q.4.h.and/or tramadol 2–5 mg/kg per os (p.o.) q.6.h. the follow-ing 7 days. All dogs were treated postoperatively with anti-biotics (amoxicillin/clavulanic acid 12.5 mg/kg q.12.h p.o.)during 5 days. Dogs were monitored daily throughout thestudy by a veterinarian to assess pain symptoms accordingto the short form of the Glasgow composite pain scale.Dogs that showed signs of pain, received tramadol and/orbuprenorphine and/or gabapentin (5 mg/kg p.o. q.12.h).Furthermore, animals were monitored daily by a veter-inarian for clinical signs of illness, neurologic deficitsand lameness.

Injected substancesIn the first study spontaneously degenerated IVDs (Pfirmanngrade 2) of the dogs were injected with a volume of30 μl of NaCl 0.9 % (sham), a bolus of CXB (7.7 μM),a CXB-loaded (7.7 μM) and an unloaded pNIPAAMMgFe-LDH hydrogel. Based on studies in cadavericspines (unpublished data, N. Willems, and B.P. Meij)a volume of 30 μl could be injected into the NP withoutsubstantial resistance. The volume of 30 μl contained7.7 μM × 10−6 M CXB, to achieve a final concentrationof 1 μM (=7.7 × 10−6 × (30 μl gel/230 μl NP volumeplus gel)) for the bolus of CXB, in the canine NP ofbeagle laboratory dogs with a mean weight of 8–9 kg andtaking into account the volume of the nucleus (200 μl)[29]. All substances were injected into the IVDs in theT12-L6 spinal segment in a randomized fashion, exceptfor the sham treatment (NaCl 0.9 %), which wasinjected into T12-T13. An interim statistical analysiswas performed after the first study to evaluate treat-ments and study design. Results were used to perform anew power analysis and to adapt the study design ofthe second study. In the second study, all substanceswere administered in a random order within each ani-mal and IVDs of the T12-S1 spinal segment wereinjected with NaCl 0.9 %, a bolus of CXB (7.7 μM),CXB-loaded (7.7 μM, 77 μM and 770 μM) hydrogels,and an unloaded pNIPAAM MgFe-LDH hydrogel. IVDsadjacent to those injected with hydrogel loaded withthe highest dose of CXB (770 μM) remained untreated.

Postmortem collection of materialsDogs were euthanized 4 weeks postinjection. First, theywere sedated with dexmedetomidine 0.04 mg/kg i.v.,followed by pentobarbital 200 mg/kg i.v. Immediatelyafter euthanasia, the vertebral column (T12-S1) washarvested by using an electric multipurpose saw (Bosch,Stuttgart, Germany). All muscles were removed and thevertebrae were transected transversely with a band saw(EXAKT tape saw, EXAKT Advanced TechnologiesGmbH, Norderstedt, Germany), resulting in nine spinal


units (endplate–IVD–endplate). These units were thentransected sagittally by using a diamond band pathologysaw (EXAKT 312 saw; EXAKT diamond cutting band0.1 mm D64; EXAKT Advanced Technologies GmbH),generating two identical parts. One half was resected witha surgical knife by removing the endplate and the vertebraattached to it on one side, and the remaining IVD tissuewas snap frozen in liquid nitrogen stored at −80 °C forbiochemical and biomolecular analyses. The other halfwas photographed (Olympus VR-340, Olympus EuropaGmbH) for macroscopic evaluation of the IVD (Thompsonscore, see below) and stored for 14 days in 50 ml of 4 %buffered formaldehyde at 4 °C for histological analyses.

Histology, COX-2 immunohistochemistry, and TUNEL assaySamples were decalcified in 35 % formic acid and 6.8 %sodium formate in a microwave oven (Milestone MicrowaveLaboratory Systems, Bergamo, Italy) overnight at 37 °C, for7 nights [30] and embedded in paraffin. Five-μm-thicksections were stained with hematoxylin and eosin andwith picrosirius red/alcian blue and evaluated accordingto a grading scheme according to Bergknut et al. [31].Histological slides were scored blinded and in randomorder by two independent investigators (NW, AT) usingan Olympus BX41 microscope (Olympus Europa GmbH).In case of doubt, samples were also scored by a board-certified veterinary pathologist (GG). All photographs ofthe macroscopy of the IVD segments were evaluated bytwo independent blinded investigators (NW, AT) ac-cording to the Thompson grading scheme, which hasbeen validated in dogs [32].Immunohistochemistry for COX-2 was performed on

5 μm sections mounted on KP-plus glass slides. Afterdeparaffinization and rehydration, sections were treatedwith Dual Endogenous Enzyme Block (Dako S2003, Dako,Carpinteria, CA, USA) for 10 min at room temperature toblock nonspecific endogenous peroxidase, followed by twowashing steps each of 5 min with Tris-buffered saline with1 % Tween 20® (TBST). Sections were treated withTris-buffered saline (TBS) bovine serum albumin (BSA)5 % solution to block nonspecific binding for 60 min atroom temperature, were carefully rinsed and subsequentlyincubated with a primary mouse anti-human monoclonalCOX-2 antibody (Cayman Chemical, Ann Arbor, MI, USA)diluted 1:50 in TBS-BSA 5 % overnight at 4 °C. The follow-ing day sections were incubated with peroxidase-labeledpolymer (Envision™ anti-mouse K4001, Dako). Antibodybinding was visualized by using diaminobenzidine (DAB;Dako). Sections were counterstained with hematoxylinsolution (Hematoxylin QS, Vector Laboratories Ltd.,Peterborough, UK) and mounted in permanent mount-ing medium.A commercial available terminal deoxynucleotidyl trans-

ferase dUTP nick-end labeling (TUNEL; Merck Millipore,

Darmstadt, Germany) assay was used according to themanufacturer’s instructions to determine apoptosis.The percentage of COX-2-positive and TUNEL-positivechondrocytes over the total number of cells was deter-mined by manual counting in the NP, and in the ventral(VAF) and dorsal AF (DAF), by two blinded independ-ent investigators (NW, SP).

Biomolecular and biochemical analysesCryosections (60 μm) of the spinal units were cut with acryostat (Leica CM1800 cryostat, Leica MicrosystemsInc., Bannockburn, IL, USA) and collected on RNAse-freeglass slides. The NP and AF tissues were separated andhalf of the slides were collected in respectively 400 μland 750 μl Ambion® KDalert™ lysis buffer solution (LifeTechnologies) in the first study, and in Complete Lysis-MEDTA-free buffer (Roche Diagnostics Nederland B.V.,Almere, The Netherlands) in the second study and storedat −80 °C until biochemical analyses were performed. Theother half was collected in 300 μl RLT buffer containing1 % β-mercaptoethanol (Qiagen, Venlo, The Netherlands)and stored at −80 °C until biomolecular analyses wereperformed.Quantitative PCR (qPCR) was performed to assess the

effects of (controlled release of ) CXB at gene expressionlevels of the NP with regards to: 1) ECM anabolism: aggre-can (ACAN), collagen type II (COL2A1), collagen type I(COL1A1); 2) ECM catabolism: a disintegrin and metallo-proteinase with thrombospondin motifs 5 (ADAMTS5),matrix metalloproteinase 13 (MMP13), tissue inhibitorof metalloproteinase 1 (TIMP1); 3) inflammation: tumornecrosis factor alpha (TNFA), interleukin-1β (IL1B),interleukin-6 (IL6) and interleukin-10 (IL10); 4) COXpathway and PGE2 synthesis: prostaglandin E synthase1 (PTGES1), prostaglandin E synthase 2 (PTGES2), cy-clooxygenase 1 (COX1), and cyclooxygenase 2 (COX2);5) notochordal markers: brachyury (T), cytokeratin-8(CK8), cytokeratin-18 (CK18); 6) the indirect effect ofCXB on Wnt signaling pathway: axin-2 (AXIN2), c-Myc(c-Myc) and cyclin-D1 (CCND1) and 7) apoptosis:caveolin-1 (CAV1), caspase 3 (CASP3), fas ligand (FasL)and Bcl-2 (BCL2). The primer pairs used for qPCR aregiven in Additional file 2.The RNeasy Fibrous Tissue Mini Kit (Qiagen, Venlo, The

Netherlands) was used to isolate total RNA. To maximizeRNA yield, the incubation period with proteinase K wasreduced to 5 min. After on-column DNase-I digestion(Qiagen RNase-free DNase kit) RNA was quantified byusing a NanoDrop 1000 spectrophotometer (Isogen LifeScience B.V., Ijsselstein, The Netherlands). cDNA wassynthesized from 20 ng total RNA in a total volume of15 μl using the iScript™ cDNA Synthesis Kit (Bio-RadLaboratories B.V., Veenendaal, The Netherlands). qPCRwas performed in duplicate using an iCycler CFX384


Touch™ thermal cycler, and IQ SYBRGreen Super mix(Bio-Rad Laboratories). All dog-specific primers weredesigned in-house using Perlprimer [33] except for MMP13[34]. Primer specificity was evaluated with BLAST, andthe designed amplicon was tested for secondary struc-tures using MFold [35]. Primers were purchased fromEurogentec, Maastricht, The Netherlands. Amplificationefficiencies ranged from 80 to 115 %. Relative expressionlevels were determined by normalizing the cycle threshold(Ct) value of each target gene by the mean Ct value ofthree reference genes, i.e., glyceraldehyde 3-phosphatedehydrogenase (GAPDH), ribosomal protein S19 (RPS19),and TATA-binding protein (TBP).To measure glycosaminoglycan (GAG) and DNA con-

tent in the NP and AF, samples in Ambion® KDalert™lysis buffer solution were homogenized in a tube rotatorO/N at 4 °C, whereas samples in Complete Lysis-MEDTA-free buffer were homogenized in a TissueLyserII (Qiagen) for 2 × 30 s at 20 Hz. The supernatant andpellet of each NP and AF were digested overnight in apapain buffer (250 μg/ml papain (Sigma-Aldrich) in 50 mMEDTA and 5 mM L-cysteine) at 60 °C. GAG content wasquantified by using a 1,9-dimethylmethylene blue assay[36]. The Quant-iT™ dsDNA Broad-Range assay kit in com-bination with a Qubit™ fluorometer (Invitrogen, Carlsbad,CA, USA) was used in accordance with the manufacturer’sinstructions to determine DNA content in the papain-digested NP and AF supernatant and pellets. DNA con-tent in the supernatants of the NP and AF were negligibleand therefore not included in the total content.PGE2 levels were measured with the same colorimetric

competitive enzyme immunoassay kit (PGE2 high sensi-tivity EIA kit, Enzo Life Sciences BVBA) that was usedfor the in vitro experiments. Both buffers that were usedto lyse tissue were validated and standards were dilutedin the same lysis buffer as the samples, which did not showstrong interference with the performance of the kit. TotalGAG content and PGE2 levels were normalized for DNAcontent in the pellet and were measured in the NP as wellas the AF.

Statistical analysesPower analyses were performed prior to both in vivostudies by using free software [51], and are described indetail in Additional file 3. PGE2/DNA in the NP was con-sidered to be the main read-out parameter. Biochemicaland biomolecular data were analyzed by using the R statis-tical software, package 2.15.2. A linear mixed-effectsmodel was used to analyze the effect of the injected treat-ments. Factors incorporated into the model as a fixedeffect were ‘treatment’ (NaCl, CXB 7.7 μM, CR, CR +7.7 μM, CR + 77 μM, CR + 770 μM), ‘tissue’ (NP and AF),and their interaction. Random effects ‘dog’ (dog 1–18)and ‘study’ (study 1 and 2) were incorporated to capture

the correlation between multiple measurements withinone dog. Residual plots and quantile–quantile (QQ)-plotswere used to check for possible violations of normality as-sumptions. In case of violation, data were logarithmicallytransformed. The Cox proportional hazards regressionmodel was used to estimate the effect of the injected treat-ments on gene expression levels. Calculations were per-formed on Ct values for each target gene and the mean Ctvalue of three reference genes was incorporated into themodel as a covariate. If proportional hazard assumptionswere violated, the ratio of the Ct values for each targetgene to the mean Ct value of the reference genes was usedfor analysis. Ct values ≥ 40 were right censored. Regressioncoefficients were estimated by the maximum likelihoodmethod. Model selection was based on the lowest Akaikeinformation criterion (AIC). Confidence intervals werecalculated and stated at the 99 % confidence level to correctfor multiple comparisons. Differences between treatmentswere considered significant if the confidence interval didnot include 0, whereas hazard ratios were consideredsignificant if the confidence interval did not include 1.

ResultsRheological properties and handling of the pNIPAAMMgFe-LDH hydrogelsAt low temperatures (22 °C) the complex modulus |G*| =10 Pa and at high temperatures (37 °C) the complexmodulus |G*| = 2 kPa. Typical mechanical properties of aNP are in the range of 7–21 kPa [38, 39]. However, thishydrogel was not intended to be used as a replacement forthe NP and therefore no load-bearing properties wereneeded. The slightly lower modulus of the hydrogel wassufficient for the purpose of controlled drug release, andfor the intradiscal injection of a small volume, without in-creasing intradiscal pressure with inherent effects onhomeostasis of the resident cells. Sterilization with gammaradiation did not have a significant effect on the rheo-logical properties of the gel (Fig. 1b). The viscous hydrogeltransitioned from a low-viscous state to a stable hydrogelstate within 10 s due to the hydrophobic interactionsbetween the isopropyl groups of pNIPAAM uponincreasing the temperature above its lower critical solu-tion temperature (LCST) 32 °C (Fig. 3a and 3b). The lowviscosity solution could easily be injected at roomtemperature via a 29G needle.

Degradation and controlled release of CXB from pNIPAAMMgFe-LDH hydrogels in vitroA cumulative release of 14 % CXB from hydrogelsloaded with 10 mg/ml CXB was shown after 31 days,whereas a cumulative release of 11 % Mg was detected,the former indicating that CXB is still present in thehydrogel and controlled release of CXB is most probablyaccomplished over more than 31 days. The release of

Fig. 3 a Schematic diagram of the formation of stable hybrid hydrogels. Positively charged layered double hydroxides (LDH) in a pNIPAAM solution(pNIPAAM chain) transit from a low-viscous state to a stable hydrogel due to hydrophobic interactions between the isopropyl groups of pNIPAAMupon increasing temperature. The celecoxib (CXB) is present in the hydrogel in small crystals, forming a depot, from which dissolution and diffusiontakes place. b At 37 °C gelation occurs within 10 s. c One month after intradiscal injection the pNIPAAM MgFe-LDH hydrogel (white arrow)is visible in the nucleus pulposus (NP; **). The annulus fibrosus (AF) is indicated with a black arrow. LDH layered double hydroxide, pNIPAAMpoly-N-isopropylacrylamide


CXB showed a similar pattern as dissolution of the MgFe-LDH particles in PBS/0.2 % Tween® (Additional file 1).An increase in the concentration of Tween® is knownto accelerate gel degradation and increase CXB solubilityand resulted in a threefold higher Mg release and a two-to threefold higher release of CXB (Additional file 1) [22].Neither an increase in the amount of LDH particles, northe charge of Mg affected the CXB release profile. Fur-thermore, a 1.5-fold increase in the cumulative release ofCXB could be detected in gels with a 6 mg/ml loadingdose of CXB compared with a 10 mg/ml loading dose.However, the absolute amount of CXB was comparable(20 % of 10 mg/ml versus 30 % of 6 mg/ml) (Additionalfile 1). Hydrogels with higher amounts of LDH parti-cles showed a lower amount of cumulative Mg release(Additional file 1). Degradation of pNIPAAM MgFe-LDHhydrogels and controlled release of CXB in vitro are illus-trated in the figures depicted in Additional file 1.Agarose–cell constructs incubated with 10 μM CXB

bolus for 2 consecutive days apparently had taken upCXB by diffusion, leading to detectable amounts of CXBin the medium up to 7 days, which then dropped to zero(Fig. 2b). CXB released from the 0.1 mg/ml loaded hydro-gels into the culture medium ranged from 1.1 to 4.2 μM(Additional file 1). CXB released from hydrogels loadedwith 1 mg/ml CXB resulted in CXB concentrations in themedium ranging from 3 to 15.9 μM. To determine the ac-tivity of the released CXB, cells in the constructs werestimulated by 10 ng/ml TNF-α, which resulted in

detectable PGE2 levels in the culture medium at day 9.Application of the 10 μM bolus of CXB during 2 consecu-tive days resulted in suppression of PGE2 levels from day9 to 11, and PGE2 levels started to increase afterward tosimilar levels as the TNF-α-stimulated constructs. In theconstructs cultured in the presence of 1 mg/ml CXB-loaded hydrogels, TNF-α-induced PGE2 production wascompletely inhibited throughout the whole culture periodof 28 days (Fig. 2c).

In vivo biocompatibility in miceSubcutaneous injection of the different hydrogels showedno adverse local or systemic effects. At 7 days postinjec-tion a fibrous capsule of varying thickness (57–76 μm)was present in all hydrogel-injected tissue samples. At theinterface between the hydrogel and this capsule mainlyneutrophils and macrophages were present, consistentwith an acute/subacute pyogranulomatous reaction (Fig. 4).At 28 days postinjection the hydrogel and fibrous capsuleswere also present. In two samples a decreased thicknessof the fibrotic capsules (35 and 40 μm) was observed,whereas in one sample a thickened capsule (236 μm) wasobserved. In all three samples macrophages constitutedthe predominant cell type in the intermediate layer, con-sistent with a granulomatous reaction. Some of these mac-rophages showed marked evidence of phagocytic activity.In two samples injected with PBS (control), a slight in-crease in macrophages was seen at day 28 compared with

Fig. 4 Representative histological images of subcutaneous injection sites in mice (hematoxylin and eosin stain). a-c show the skin and subcutisof a control animal at 7 days revealing the epidermis (arrow a), hair follicles (arrowhead a) and striated muscle (panniculus carnosus; asterisk aand b). The panniculus carnosus (asterisk c) and an occasional mastcell (arrowhead c) are visible in the subcutis. d-f show the skin 7 days after injectionof the hydrogel. The hydrogel is visible as a grey, granular substance (white arrow e) positioned below the panniculus carnosus (arrowhead e)surrounded by a capsule of loosely arranged fibroblasts (asterisk e). Multifocally infiltrates of eosinophils (white arrow f), neutrophils (arrowhead f) andmacrophages (black arrow f) separating the hydrogel from the fibrous capsule. g-i show the skin of a control animal at 28 days after injection withoutsignificant pathological changes. j-l depict the histological changes in the subcutis 28 days after injection of the hydrogel with the epidermis andpanniculus carnosus indicated by an arrow and arrowhead respectively in j. k shows a more compact fibrous capsule (arrow) compared to the loosecapsule seen after 7 days postinjection. The cellular reaction directly surrounding the hydrogel show a more granulomatous nature indicated by thepresence of macrophages often containing brown pigment l



day 7. Giant cells, necrosis, neovascularization, and fattyinfiltration were not observed at either time point.

Intradiscal application and controlled release of CXB-loaded hydrogels in laboratory beagle dogsSurgical follow-upBefore surgery a total of 162 IVDs were graded on MRimages. A total of 9/162 IVDs were assigned a grade Iaccording to the Pfirmann system, whereas 153/162 IVDswere assigned a grade II, of which 88 were injected in thisstudy. Six out of nine dogs in the first study were ambu-lant the day after the injections of the test substances inthe IVD and showed a slight reduction in spinal reflexesthat recovered within the following 7 days. Three dogsalso showed reduced weight bearing of the left hind limb,and received pain medication for a longer period of time.Two dogs that received pain medication for 7 more days,recovered completely. In one of these dogs slight dehis-cence of the wound was detected and antibiotics weregiven for a total of 14 days. In one dog the reduction inspinal reflexes and weight bearing of the left hind limbpersisted and this dog was also treated with gabapentin 5mg/kg p.o. q.12.h. All dogs in the second study showeduneventful recovery from surgery, were ambulant thenext day and showed minor reductions in spinal reflexesthat recovered within 7 days.

IVD integrityPostmortem, Thompson score grade II was assigned to87/88 IVDs; a grade III was assigned to 1/88 IVDs,which had been injected with the unloaded hydrogel. In17/52 IVDs injected with the hydrogel, the tan-coloredhydrogel was visible in the mid-sagittal sections of theNP (Fig. 3c). Histological evaluation of the total of 88IVDs was performed as per the grading scheme accord-ing to Bergknut et al. [31]. Scores ranged from 4 to 14.The median histological grade in the first study was 9.5(4–13) and in the second study 11 (8–14). No significantdifferences were found between the injected treatments.In one of the IVDs injected with the empty hydrogel(level L6-L7), fibrotic tissue was present in the dorsalAF; in another IVD injected with the empty hydrogel(level L7-S1), the central parts of both sides of the end-plates (EPs) were very irregular and clusters of chondro-cytes were present in the dorsal AF at both sides of theAF-EP interface. In one of the IVDs injected with NaCl(level T12-T13), acellular material was detected in theventral AF. At macroscopic examination slight bulgingof the ventral AF was noticed.Relative gene expression levels of BCL2, a regulatory

gene of cell death (apoptosis) were significantly down-regulated in the CXB-loaded hydrogel compared with thesham (HR = 8.28, CI 99 % 1.45–47.19) (Fig. 5a). However,gene expression levels of other apoptotic markers, i.e.,

CAV1, CASP3, and FasL showed no signs of increasedapoptosis in any of the treatments. Percentages ofTUNEL-positive cells per total cell count showed nodifferences between treatments either (median 0 %,range 0–82 %). Furthermore, there were no significantdifferences in gene expression levels of notochordal cellmarkers T, CK8, CK18, nor in levels of AXIN2, c-Myc,and CCND1, associated with the Wnt pathway betweenthe treatments.

Extracellular matrix metabolismRelative gene expression of the catabolic gene ADAMTS5was significantly upregulated in the NP samples treatedwith the CXB bolus (HR = 10.35, CI 99 % 1.74–61.57) andthe CXB-loaded hydrogel (HR = 10.66, CI 99 % 1.70–66.67) (Fig. 5b) compared with the unloaded hydrogel.Gene expression levels of other catabolic (MMP13) andanti-catabolic (TIMP1) genes were not significantly dif-ferent between treatments (Fig. 5b). However, genes as-sociated with extracellular matrix (ECM) components,i.e., ACAN, COL2A1, COL1A1 did not significantly dif-fer between treatments (Fig. 5c). These findings wereconsistent with normalized GAG content (GAG/DNA)in the NP as well as the AF, which did not significantlydiffer between treatments 4 weeks postinjection. GAG/DNA levels were significantly higher in the NP than inthe AF in all treatments (M = 0.58, SD = 0.03, CI 99 %0.50–0.66) (Fig. 6b).

COX pathway and PGE2 levelsIn all CXB-loaded hydrogels, a decrease in PGE2/DNAlevels in the NP was detected relative to the NaCl-injected NPs, with a maximum reduction of 35 % for the77 μM CXB-loaded hydrogel. However, PGE2/DNA levelsin the NP as well as the AF showed high standard devi-ations and were not significantly different between thetreatments 4 weeks postinjection. Gene expression levelsof genes involved in PGE2 biosynthesis, i.e., PTGES1,PTGES2, COX1, and COX2, showed no significant dif-ferences between treatments either. Relative gene expres-sion levels of genes associated with inflammation, i.e.,TNFA, IL1B, IL6, and IL10 were below the detection levelfor all conditions. Cells expressing positive immunohisto-chemical COX-2 staining were detected in only three sitesout of a total of 162 (54 IVD levels), Percentages of COX-2-positive cells in the NP and VAF of two IVDs injectedwith 770 μM CXB-loaded hydrogel were low, 0.5 % and2 % respectively, and 0.6 % in an IVD injected withunloaded hydrogel. Regardless of the treatment, PGE2/DNA levels in the NP were significantly lower than thosedetected in the AF (M = −0.64, SD = 0.08, CI 99 % −0.85to −0.43) (Fig. 6c).

Fig. 5 a-c. Relative gene expression levels per treatment. The NaCl (sham) treatment in nucleus pulposus (NP) tissue is set at 1. a Relative geneexpression levels of ADAMTS5 were significantly upregulated in the NP samples treated with the celecoxib (CXB) bolus (CXB 7.7 μM) and theCXB-loaded hydrogel (CR + CXB 7.7 μM) compared with the unloaded hydrogel (CR). b Gene expression levels ACAN, COL2A1, COL1A1 did notsignificantly differ between treatments. c BCL2 expression levels were significantly downregulated in the CXB-loaded hydrogel compared withthe sham (NaCl). Data are expressed as n-fold changes ± standard deviation. d and e Representative histological images of early degeneratedcanine NP injected with NaCl (d) or a 770 μM celecoxib (CXB)-loaded pNIPAAM MgFe hydrogel (e) stained with a COX-2 antibody andcounterstained with hematoxylin. No significant differences were found between the injected treatments. Chondrocyte-like cell (asterisk) density isincreased, and small size clones are present (arrow), indicative of early IVD degeneration. None of the cells in these NPs demonstrated positivestaining for the COX-2 antibody. Treatments: NaCl = NaCl (sham); CXB 7.7 μM = celecoxib bolus; CR = unloaded pNIPAAM MgFe-LDH hydrogel;CR + CXB 7.7 μM = pNIPAAM MgFe-LDH hydrogel loaded with 7.7 μM CXB). *Indicates significant difference at a 99 % confidence level. ACANaggrecan, ADAMTS5 a disintegrin and metalloproteinase with thrombospondin motifs 5, BCL2 B-cell lymphoma 2, COL1A1 collagen type 1alpha 1, COL2A1 collagen type 2 alpha 1, COX cyclooxygenase, IVD intervertebral disc, pNIPAAM poly-N-isopropylacrylamide


DiscussionTo the authors’ knowledge this is the first study that de-scribes biocompatibility and safe intradiscal applicationof a thermoreversible pNIPAAM MgFe-LDH hydrogelin vivo. The hydrogel was successfully employed as a ve-hicle for the delivery of a COX-2 inhibitor into the IVD.The selective COX-2 inhibitor celecoxib (CXB) was se-lected, as this drug is commonly used to alleviate painsymptoms associated with degenerative IVD conditions.

In vitro controlled release of CXB results into sustainedsuppression of PGE2 levels associated with inflammationIn vitro, degradation behavior of the thermoresponsivehydrogel was comparable for different loading dosagesof CXB. Although an increase in the cumulative releaseof CXB as a percentage of the total amount loaded wasdetected in gels with a lower loading dose, the absoluteamounts measured in the medium were comparable, dem-onstrating a CXB solubility-dependent release. Hydrogels

with higher amounts of LDH particles showed a lower cu-mulative Mg release, most likely due to an increase in an-ionic exchange with the medium [40]. Neither an increasein the amount of LDH particles, nor changes in the chargeof Mg affected the CXB release profiles. In vitro, stimula-tion of three-dimensional chondrocyte constructs with 10ng/ml TNF-α, resulted in increased PGE2 levels in the cul-ture medium. Furthermore, in this in vitro model, weconfirmed the controlled release of CXB into themedium from 0.1 mg/ml and 1 mg/ml CXB-loaded hydro-gels, respectively. Unfortunately, CXB values measured bythe UPLC method showed high variances most probablydue to fact that the values were measured in the lower re-gion of the UPLC detection method [23]. TNF-α-inducedPGE2 production was completely inhibited by 1 mg/mlCXB-loaded pNIPAAM MgFe-LDH hydrogels, therebyproving the efficacy of the CXB controlled release system tosuppress in vitro PGE2 production for a prolonged period incontrast to the short-lived effect of a bolus of CXB.

Fig. 6 Representative macroscopic and histopathologic image stainedwith alcian blue/picrosirius red of an IVD treated with NaCl (a) andGAG and PGE2 levels normalized for DNA in nucleus pulposus (NP)and annulus fibrosus (AF) tissue (b-c). The NP in figure (a) has abulging aspect due to the processing method. b GAG/DNA levelswere significantly higher in the NP than in the AF in all treatments.c In all treatments PGE2/DNA levels were significantly lower in theNP than those in the AF. Data are expressed as mean values ± standarddeviation. Treatments: NaCl = NaCl (sham); CXB 7.7 μM = celecoxibbolus; CR = unloaded pNIPAAM MgFe-LDH hydrogel; CR + CXB7.7 μM = pNIPAAM MgFe-LDH hydrogel loaded with 7.7 μM CXB;CR + 77 μM = pNIPAAM MgFe-LDH hydrogel loaded with 77 μMCXB; CR + CXB 770 μM = pNIPAAM MgFe-LDH hydrogel loadedwith 770 μM CXB). *Indicates significant difference at a 99 % confidencelevel. CXB celecoxib, GAG glycosaminoglycan, IVD intervertebraldisc, LDH layered double hydroxide, PGE2 prostaglandin E2, pNIPAAMpoly-N-isopropylacrylamide


Safe subcutaneous and intradiscal application of thethermoreversible pNIPAAM MgFe-LDH hydrogelIn vivo, subcutaneous injection of the hydrogel did not re-sult in local or systemic adverse effects; histology showed

a moderately irritant reaction with a shift from a pyogra-nulomatous reaction into a granulomatous reaction withformation of a fibrous capsule, consistent with a foreignbody reaction to biomaterials [41]. Based on these clinicaland histological findings, in combination with the avas-cular nature of the IVD [42, 43], we concluded that thehydrogel would be well tolerated when applied intradis-cally. Indeed, safe intradiscal injection of thermorespon-sive CXB-loaded and unloaded pNIPAAM MgFe-LDHhydrogels was demonstrated in a large animal model, i.e.,chondrodystrophic dogs, with naturally occurring IVDdegeneration. This was corroborated by IVD histology,further determination of the anabolic/catabolic state ofthe ECM and cell viability by means of gene expressionand biochemical analyses. The presence of the hydrogelin the NP, in contrast to the subcutaneous location, wasnot accompanied by a foreign body reaction. Histologicalfindings in the majority of the canine IVDs were unre-markable. Notably, in one IVD injected with unloadedhydrogel, fibrous tissue in the AF was detected, whereasin another one, irregular EPs and clustering of chondro-cytes in the dorsal AF were present. In one of the IVDsinjected with NaCl, acellular material and bulging of theventral AF were noted. These irregularities cannot beattributed with certainty to the injections [18], and mayalso reflect spontaneous progression of IVD degener-ation. This study was conducted in mildly degeneratedIVDs without fissures in the AF and we demonstratedsafe intradiscal injection of 30 μl hydrogel through a29G needle. However, we did not study the ability ofthe pNIPAAM MgFe-LDH hydrogel to form an inter-face with NP tissue and therefore we cannot excludethe risk of extrusion of the biomaterial in severelydegenerated IVDs that contain annular fissures.Viability of the resident NP cells of the NP was not

affected by the in vivo intradiscal application of thepNIPAAM MgFe-LDH hydrogel loaded with CXB overa wide dose range. Relative gene expression levels of theanti-apoptotic gene BCL2 were significantly downregulatedin the NPs of IVDs injected with the CXB-loaded hydrogelcompared with the NaCl-injected IVDs. In numerous can-cer cell lines overexpressing COX-2, CXB has been shownto activate the intrinsic apoptotic pathway [44, 45]. How-ever, in our study, gene expression levels of other apoptoticmarkers, i.e., CAV1, CASP3, and FasL, together with resultsof the TUNEL assay showed no evidence that apoptosiswas affected by the intradiscal application of (un)loadedpNIPAAM MgFe-LDH hydrogel in vivo. Altogether thisindicates that the differences of BCL2 on gene expres-sion level may not be of biological relevance. In line withthe aforementioned, expression levels for notochordal cellmarkers T, CK8 and CK18 did not differ between groupsindicating that neither the injection nor the (un)loadedbiomaterial had an adverse effect on IVD health.


Safe intradiscal application of the thermoreversiblepNIPAAM MgFe-LDH hydrogel loaded with a wide rangeof CXB dosagesThe pNIPAAM MgFe-LDH hydrogel loaded with CXBover a wide dose range also appeared to be biocompatibleand safe for intradiscal application at a biomolecular andbiochemical level. Intradiscal injection was not associatedwith IVD herniation, or with progression of degener-ation. Overall, there were no significant differences be-tween treatments on expression levels of other catabolic(MMP13) and anti-catabolic (TIMP1) genes, genes of extra-cellular matrix (ECM) components, i.e., ACAN, COL2A1,COL1A1, and the GAG/DNA content. Furthermore, geneexpression levels of ADAMTS5 were significantly upregu-lated in NPs treated with the CXB bolus and the CXB-loaded hydrogel, independent of the dose, compared withNPs treated with the unloaded hydrogel, suggestive of acatabolic effect of CXB. In line with this, ADAMTS5 geneexpression levels did not significantly differ between IVDsinjected with unloaded hydrogels and NaCl. Our findingsdo not correspond with upregulated levels of ADAMTS4reported in bovine NPCs cultured in hyaluronan-based(HA)-pNIPAAM hydrogels compared with NPCs culturedin alginate beads. Nevertheless, these results should becompared to our results with care, as the relevant fold-change in gene expression was limited to a twofold dif-ference in our study, and the experimental environment(i.e., in vitro versus in vivo), as well as the tissue andthe composition of the hydrogel differed [46]. Altogether,the pNIPAAM MgFe-LDH hydrogel is biocompatibleand can be safely injected in the IVD without affectingthe IVD health based on the overall results of histologicalscores, immunohistochemical indices, gene expressionprofiles, and biochemical analyses.

Controlled release of CXB in mildly degenerated IVDs hada limited effect on PGE2 levelsOver the period of 28 days and there was no dose-dependent effect in a wide range of CXB concentrations.Due to technical limitations we were not able to determinethe CXB tissue levels in vivo and hence cannot elaborateon the in vivo release profile of the loaded hydrogels specif-ically in the matrix-rich NP environment. As CXB has arelatively high protein-binding capacity, the protein-richNP might have enhanced CXB solubility, hence an in-creased release of CXB from the hydrogel [47]. Given thatthe only available reports on the effect of CXB on PGE2production by IVD cells are from in vitro experiments,the discussion of the results is further limited by differ-ences between cell responses in an in vivo and anin vitro situation. Furthermore, PGE2 levels detected inthese in vitro experiments are expressed in PGE2/mlmedium or PGE2/total protein, and in none of them inPGE2/DNA [6, 7, 9, 10, 48]. When correcting the PGE2

levels for protein content in the NP (Additional file 4),the results are consistent with findings in 7-day culturesof human grade 3 degenerated tissue, which showed noinhibition of PGE2 production in the presence of 1 μMCXB [48]. Although we cannot rule out a suboptimaleffect of the CXB-loaded hydrogels in the intradiscal envir-onment, the lack of PGE2 inhibition by CXB in the currentstudy may have been attributable to the constitutive activ-ity of COX-1 and the absence of inflammation-inducedCOX-2 activity. The low percentage of COX-2-positivecells detected at the immunohistochemical level supportsthis theory. Furthermore, in patients with degenerativejoint cartilage, treatment with CXB for a period of 28days has been shown to have a beneficial effect on GAGturnover, mainly due to COX-2 inhibition. GAG/DNAwas not significantly different for any of the treatments,which may also be associated with the absence of COX-2activity in this model. Insufficient statistical power becauseof the modest sample size (N = 18) may have played a rolein limiting the significance of the statistical comparisonsconducted (Additional file 3).Although MR images, macroscopic, and histologic find-

ings were consistent with mild degeneration, this phase inthe degenerative cascade may actually not be associatedwith increased PGE2 levels, suggesting that inflammationoccurs at a later time point, i.e., when disc protrusionand/or clinical signs are present. Incorporation of generalCOX inhibitors into the hydrogel may provide clarity onthis.Interestingly, regardless of the treatment condition,

PGE2/DNA levels measured in the AF in our study weresignificantly higher than the levels detected in the NPfor all conditions, while GAG/DNA levels in the AF weresignificantly lower than levels in the NP, in accordancewith previous studies [5, 20]. In contrast, Miyamoto et al.described similar PGE2/ml levels in control NP and AFcell cultures, but showed a higher production of PGE2 byAF cells compared with NP cells in response to a cyclicmechanical load [49].Inflammation is thought to play an important role in

the process of IVD degeneration and disease [25]. Ad-dressing the origin of inflammatory responses and hencean earlier stage of the degenerative cascade, may be moresuccessful in suppressing inflammation and restoring IVDhomeostasis. Emerging treatment strategies aiming at con-trolled release of anti-inflammatory medication that targetproinflammatory mediators, e.g., TNF-α and PGE2 levelsin the IVD, will need to focus on IVDs in human and/orveterinary patients with clinical signs of IVD disease[16, 25, 50]. PGE2 levels elevated in the course of a trueinflammatory response in NP and/or AF tissue may beeffectively decreased by selective COX-2 inhibitors, suchas CXB. However, CXB release and bioavailability in theintradiscal environment should be understood in more


detail. If these issues are solved, pNIPAAM MgFe-LDHhydrogels loaded with CXB might be a promising long-term treatment in patients with herniated discs andchronic low back pain that can be minimally invasiveinjected into the diseased IVD assisted by fluoroscopyor computer tomography.

ConclusionsWe have demonstrated the biocompatibility and safetyof a pNIPAAM MgFe-LDH hydrogel via subcutaneousapplication in mice and via intradiscal administration ina large animal model. Furthermore, we have shown itscapability for controlled delivery of the anti-inflammatorydrug CXB, resulting in suppressed PGE2 levels in a TNF-α-stimulated three-dimensional tissue-engineered modelsystem for up to 28 days. The controlled release of CXBfrom this hydrogel resulted in limited inhibition of PGE2production in a large animal model with spontaneous IVDdegeneration. This may be due to the stage of degener-ation rather than the efficacy of the controlled releasesystem.

Additional files

Additional file 1: Synthesis and degradation of poly-N-isopropylacrylamide (pNIPAAM) MgFe-LDH hydrogels and controlledrelease of CXB in vitro. (DOCX 89 kb)

Additional file 2: Primers used for qPCR. (DOCX 19 kb)

Additional file 3: Power analysis of the studies in laboratory beagledogs. (DOCX 16 kb)

Additional file 4: PGE2 normalized for total protein. (TIFF 15443 kb)

AbbreviationsACAN: aggrecan; ADAMTS5: a disintegrin and metalloproteinase withthrombospondin motifs 5; AF: annulus fibrosus; AIC: Akaike informationcriterion; BCL-2: B-cell lymphoma 2; BSA: bovine serum albumin;c.r.i.: continuous rate infusion; CASP3: caspase 3; CAV1: caveolin-1;CCND1: cyclin-D1; CK: cytokeratin; COL1A1: collagen type 1 alpha 1;COL2A1: collagen type 2 alpha 1; COX: cyclooxygenase; Ct: cycle threshold;CXB: celecoxib; DEC: Animal Experiments Committee (Dutch:Dierexperimentencommissie); ECM: extracellular matrix; FasL: fas ligand;FSE: fast spin echo; G*: complex shear modulus; G’: storage modulus; G”: lossmodulus; GAG: glycosaminoglycan; GAPDH: glyceraldehyde-3-phosphatedehydrogenase; hgDMEM: high-glucose Dulbecco’s modified Eagle’smedium; i.m.: intramuscular; i.p.: intraperitoneal; i.v.: intravenous;IL: interleukin; IVD: intervertebral disc; LCST: lower critical solution temperature;LDH: layered double hydroxide; MMP13: matrix metalloproteinase 13;MR: magnetic resonance; NP: nucleus pulposus; p.o.: per os; PBS: phosphate-buffered saline; PGE2: prostaglandin E2; PGH2: prostaglandin H2;pNIPAAM: poly-N-isopropylacrylamide; PTGES: prostaglandin E synthase;q: quaque; qPCR: quantitative polymerase chain reaction; RPS19: ribosomalprotein S19; T: brachyury; TBP: TATA-binding protein; TBS: Tris-bufferedsaline; TBST: Tris-buffered saline with 1 % Tween 20; TE: echo time;TIMP1: tissue inhibitor of metalloproteinase 1; TNF-α: tumor necrosis factoralpha; TR: repetition time; TUNEL: terminal deoxynucleotidyl transferasedUTP nick-end labeling; UPLC: ultra-performance liquid chromatography.

Competing interestsThe author(s) declare that they have no competing interests. The fundingbody of these studies did not have an influence on the study design andthe data interpretation.

Authors’ contributionsNW drafted the manuscript, and participated in the study design, acquisition,analysis and interpretation of data. ML, EC, and NP prepared the hydrogelsand contributed to the in vitro study design, acquisition and interpretationof the in vitro data. HY and HK participated in the acquisition and interpretationof the in vivo data in mice. AT, SP, and FR contributed to the acquisition andinterpretation of the in vivo data in dogs. GG participated in the histologicalexamination of the intervertebral discs. WD and LC participated in the studydesign, and interpretation of data. BM and MT participated in the study design,acquisition and interpretation of data. All authors critically revised and approvedthe final manuscript.

AcknowledgementsThis research forms part of the Project P2.01 IDiDAS of the research programof the BioMedical Materials Institute, co-funded by the Dutch Ministry ofEconomic Affairs, Agriculture and Innovation. The financial contribution ofthe Dutch Arthritis Foundation is gratefully acknowledged (IDiDAS and LLP22).We would like to acknowledge TNO for supplying us with the biomaterial, HansVernooij and Paul Westers for their help with the statistical analyses, and RuudLicht for his technical assistance in the laboratory.

Author details1Department of Clinical Sciences of Companion Animals, Faculty ofVeterinary Medicine, Utrecht University, Yalelaan 108, Utrecht 3584CM, TheNetherlands. 2Department of Orthopedics, University Medical Center,Heidelberglaan 100, Utrecht 3584CX, The Netherlands. 3Department ofMaterials Technology, TNO, De Rondom 1, Eindhoven 5612AP, TheNetherlands. 4Department of Pathobiology, Faculty of Veterinary Medicine,Utrecht University, Yalelaan 1, Utrecht 3508TD, The Netherlands.

Received: 4 June 2015 Accepted: 24 July 2015

References1. de Schepper EI, Damen J, van Meurs JB, Ginai AZ, Popham M, Hofman A, et al.

The association between lumbar disc degeneration and low back pain:the influence of age, gender, and individual radiographic features. Spine(Phila Pa 1976). 2010;35:531–6.

2. Adams MA, Dolan P, McNally DS. The internal mechanical functioning ofintervertebral discs and articular cartilage, and its relevance to matrix biology.Matrix Biol. 2009;28:384–9.

3. Hirsch C, Schajowicz F. Studies on structural changes in the lumbar annulusfibrosus. Acta Orthop Scand. 1953;22:184–231.

4. Seguin CA, Pilliar RM, Roughley PJ, Kandel RA. Tumor necrosis factor-alphamodulates matrix production and catabolism in nucleus pulposus tissue.Spine (Phila Pa 1976). 2005;30:1940–8.

5. Roughley PJ. Biology of intervertebral disc aging and degeneration: involvementof the extracellular matrix. Spine (Phila Pa 1976). 2004;29:2691–9.

6. Ahn SH, Cho YW, Ahn MW, Jang SH, Sohn YK, Kim HS. mRNA expression ofcytokines and chemokines in herniated lumbar intervertebral discs. Spine(Phila Pa 1976). 2002;27:911–7.

7. Kang JD, Georgescu HI, McIntyre-Larkin L, Stefanovic-Racic M, Donaldson3rd WF, Evans CH. Herniated lumbar intervertebral discs spontaneouslyproduce matrix metalloproteinases, nitric oxide, interleukin-6, andprostaglandin E2. Spine (Phila Pa 1976). 1996;21:271–7.

8. Walter BA, Purmessur D, Likhitpanichkul M, Weinberg A, Cho SK, Qureshi SA,et al. Inflammatory kinetics and efficacy of anti-inflammatory treatments onhuman nucleus pulposus cells. Spine (Phila Pa 1976). 2015;40:955-963.

9. Le Maitre CL, Freemont AJ, Hoyland JA. The role of interleukin-1 in thepathogenesis of human intervertebral disc degeneration. Arthritis Res Ther.2005;7:R732–45.

10. Vo NV, Sowa GA, Kang JD, Seidel C, Studer RK. Prostaglandin E2 andprostaglandin F2alpha differentially modulate matrix metabolism of humannucleus pulposus cells. J Orthop Res. 2010;28:1259–66.

11. Park JY, Pillinger MH, Abramson SB. Prostaglandin E2 synthesis andsecretion: the role of PGE2 synthases. Clin Immunol. 2006;119:229–40.

12. Brock TG, McNish RW, Peters-Golden M. Arachidonic acid is preferentiallymetabolized by cyclooxygenase-2 to prostacyclin and prostaglandin E2. JBiol Chem. 1999;274:11660–6.

13. Birbara CA, Puopolo AD, Munoz DR, Sheldon EA, Mangione A, Bohidar NR, et al.Treatment of chronic low back pain with etoricoxib, a new cyclo-oxygenase-2

http://arthritis-research.com/content/supplementary/s13075-015-0727-x-s1.docxhttp://arthritis-research.com/content/supplementary/s13075-015-0727-x-s2.docxhttp://arthritis-research.com/content/supplementary/s13075-015-0727-x-s3.docxhttp://arthritis-research.com/content/supplementary/s13075-015-0727-x-s4.tiff


selective inhibitor: improvement in pain and disability–a randomized, placebo-controlled, 3-month trial. J Pain. 2003;4:307–15.

14. Pallay RM, Seger W, Adler JL, Ettlinger RE, Quaidoo EA, Lipetz R, et al.Etoricoxib reduced pain and disability and improved quality of life inpatients with chronic low back pain: a 3 month, randomized, controlledtrial. Scand J Rheumatol. 2004;33:257–66.

15. Hoffman AS. Hydrogels for biomedical applications. Adv Drug Deliv Rev.2002;54:3–12.

16. Likhitpanichkul M, Kim Y, Torre OM, See E, Kazezian Z, Pandit A, et al.Fibrin-genipin annulus fibrosus sealant as a delivery system for anti-TNFalphadrug. Spine J. 2015. doi: 10.1016/j.spinee.2015.04.026.

17. Bikram M, West JL. Thermo-responsive systems for controlled drug delivery.Expert Opin Drug Deliv. 2008;5:1077–91.

18. Elliott DM, Yerramalli CS, Beckstein JC, Boxberger JI, Johannessen W, Vresilovic EJ.The effect of relative needle diameter in puncture and sham injection animalmodels of degeneration. Spine (Phila Pa 1976). 2008;33:588–96.

19. Alini M, Eisenstein SM, Ito K, Little C, Kettler AA, Masuda K, et al. Are animalmodels useful for studying human disc disorders/degeneration? Eur Spine J.2008;17:2–19.

20. Bergknut N, Rutges JP, Kranenburg HC, Smolders LA, Hagman R, Smidt HJ,et al. The dog as an animal model for intervertebral disc degeneration?Spine (Phila Pa 1976). 2012;37:351–8.

21. Yang HY, van Ee RJ, Timmer K, Craenmehr EG, Huang JH, Oner C et al.A novel injectable thermoresponsive and cytocompatible gel of poly(N-isopropylacrylamide) with layered double hydroxides facilitates siRNAdelivery into chondrocytes in 3D culture. Acta Biomater. 2015;23:214-228.

22. Homar M, Ubrich N, El Ghazouani F, Kristl J, Kerc J, Maincent P. Influence ofpolymers on the bioavailability of microencapsulated celecoxib. JMicroencapsul. 2007;24:621–33.

23. Petit A, Sandker M, Muller B, Meyboom R, van Midwoud P, Bruin P, et al.Release behavior and intra-articular biocompatibility of celecoxib-loadedacetyl-capped PCLA-PEG-PCLA thermogels. Biomaterials. 2014;35:7919–28.

24. Goldring SR, Goldring MB. The role of cytokines in cartilage matrixdegeneration in osteoarthritis. Clin Orthop Relat Res. 2004; 427Suppl:S27-36.

25. Risbud MV, Shapiro IM. Role of cytokines in intervertebral disc degeneration:pain and disc content. Nat Rev Rheumatol. 2014;10:44–56.

26. Mastbergen SC, Jansen NW, Bijlsma JW, Lafeber FP. Differential direct effectsof cyclo-oxygenase-1/2 inhibition on proteoglycan turnover of humanosteoarthritic cartilage: an in vitro study. Arthritis Res Ther. 2006;8:R2.

27. Bergknut N, Auriemma E, Wijsman S, Voorhout G, Hagman R, Lagerstedt AS,et al. Evaluation of intervertebral disk degeneration in chondrodystrophicand nonchondrodystrophic dogs by use of Pfirrmann grading of imagesobtained with low-field magnetic resonance imaging. Am J Vet Res.2011;72:893–8.

28. Willems N, Bach FC, Plomp SG, van Rijen MH, Wolfswinkel J, Grinwis GC,et al. Intradiscal application of rhBMP-7 does not induce regeneration in acanine model of spontaneous intervertebral disc degeneration. Arthritis ResTher. 2015;17:137. doi:10.1186/s13075-015-0625-2.

29. Kranenburg HC, Meij BP, Onis D, van der Veen AJ, Saralidze K, Smolders LA,et al. Design, synthesis, imaging, and biomechanics of a softness-gradienthydrogel nucleus pulposus prosthesis in a canine lumbar spine model. JBiomed Mater Res B Appl Biomater. 2012;100:2148–55.

30. Kristensen HK. An improved method of decalcification. Stain Technol.1948;23:151–4.

31. Bergknut N, Meij BP, Hagman R, de Nies KS, Rutges JP, Smolders LA, et al.Intervertebral disc disease in dogs - part 1: a new histological gradingscheme for classification of intervertebral disc degeneration in dogs. Vet J.2013;195:156–63.

32. Bergknut N, Grinwis G, Pickee E, Auriemma E, Lagerstedt AS, Hagman R,et al. Reliability of macroscopic grading of intervertebral disk degenerationin dogs by use of the Thompson system and comparison with low-fieldmagnetic resonance imaging findings. Am J Vet Res. 2011;72:899–904.

33. Marshall OJ. PerlPrimer: cross-platform, graphical primer design for standard,bisulphite and real-time PCR. Bioinformatics. 2004;20:2471–2.

34. Muir P, Danova NA, Argyle DJ, Manley PA, Hao Z. Collagenolytic proteaseexpression in cranial cruciate ligament and stifle synovial fluid in dogs withcranial cruciate ligament rupture. Vet Surg. 2005;34:482–90.

35. Zuker M. Mfold web server for nucleic acid folding and hybridizationprediction. Nucleic Acids Res. 2003;31:3406–15.

36. Farndale RW, Sayers CA, Barrett AJ. A direct spectrophotometric microassayfor sulfated glycosaminoglycans in cartilage cultures. Connect Tissue Res.1982;9:247–8.

37. Iatridis JC, Weidenbaum M, Setton LA, Mow VC. Is the nucleus pulposus asolid or a fluid? Mechanical behaviors of the nucleus pulposus of the humanintervertebral disc. Spine (Phila Pa 1976). 1996;21:1174–84.

38. Iatridis JC, Setton LA, Weidenbaum M, Mow VC. The viscoelastic behavior ofthe non-degenerate human lumbar nucleus pulposus in shear. J Biomech.1997;30:1005–13.

39. Rives V, Del Arco M, Martin C. Layered double hydroxides as drug carriersand for controlled release of non-steroidal antiinflammatory drugs (NSAIDs):a review. J Control Release. 2013;169:28–39.

40. Anderson JM, Rodriguez A, Chang DT. Foreign body reaction tobiomaterials. Semin Immunol. 2008;20:86–100.

41. Bibby SR, Jones DA, Lee RB, Yu J, Urban JPG. The pathophysiology of theintervertebral disc. Joint Bone Spine. 2001;68:537–42.

42. Das DB, Welling A, Urban JP, Boubriak OA. Solute transport in intervertebraldisc: experiments and finite element modeling. Ann N Y Acad Sci.2009;1161:44–61.

43. Catalano A, Graciotti L, Rinaldi L, Raffaelli G, Rodilossi S, Betta P, et al.Preclinical evaluation of the nonsteroidal anti-inflammatory agent celecoxibon malignant mesothelioma chemoprevention. Int J Cancer. 2004;109:322–8.

44. Grosch S, Maier TJ, Schiffmann S, Geisslinger G. Cyclooxygenase-2(COX-2)-independent anticarcinogenic effects of selective COX-2inhibitors. J Natl Cancer Inst. 2006;98:736–47.

45. Peroglio M, Grad S, Mortisen D, Sprecher CM, Illien-Junger S, Alini M, et al.Injectable thermoreversible hyaluronan-based hydrogels for nucleuspulposus cell encapsulation. Eur Spine J. 2012;21:S839–49.

46. Paulson SK, Vaughn MB, Jessen SM, Lawal Y, Gresk CJ, Yan B, et al.Pharmacokinetics of celecoxib after oral administration in dogs andhumans: effect of food and site of absorption. J Pharmacol Exp Ther.2001;297:638–45.

47. van Dijk B, Potier E, Licht R, Creemers L, Ito K. The effect of a cyclooxygenase 2inhibitor on early degenerated human nucleus pulposus explants. GlobalSpine J. 2014;4:33–40.

48. Miyamoto H, Doita M, Nishida K, Yamamoto T, Sumi M, Kurosaka M. Effectsof cyclic mechanical stress on the production of inflammatory agents bynucleus pulposus and anulus fibrosus derived cells in vitro. Spine (Phila Pa1976). 2006;31:4–9.

49. Kadow T, Sowa G, Vo N, Kang JD. Molecular basis of intervertebral discdegeneration and herniations: what are the important translational questions?Clin Orthop Relat Res. 2015;473:1903–12.

50. Lenth RV. Java applets for power and sample size (computer software).2006–9. Available at http://www.stat.uiowa.edu/~rlenth/Power

51. R Development Core Team. R: a language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria. 2011.Available at http://www.r-project.org/

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit

http://dx.doi.org/10.1016/j.spinee.2015.04.026http://dx.doi.org/10.1186/s13075-015-0625-2http://www.stat.uiowa.edu/~rlenth/Powerhttp://www.r-project.org/

AbstractIntroductionMethodsResultsConclusions

IntroductionMaterials and methodsSynthesis and preparation of pNIPAAM MgFe-LDH hydrogelsRheological analysis of the pNIPAAM MgFe-LDH hydrogelsDegradation and release behavior of CXB-loaded pNIPAAM MgFe-LDH hydrogelsControlled release of CXB invitroIn vivo biocompatibility in mice after subcutaneous implantationIntradiscal application of CXB-loaded pNIPAAM MgFe-LDH hydrogels in laboratory beagle dogsSurgical procedureInjected substancesPostmortem collection of materialsHistology, COX-2 immunohistochemistry, and TUNEL assayBiomolecular and biochemical analysesStatistical analyses

ResultsRheological properties and handling of the pNIPAAM MgFe-LDH hydrogelsDegradation and controlled release of CXB from pNIPAAM MgFe-LDH hydrogels invitroIn vivo biocompatibility in miceIntradiscal application and controlled release of CXB-loaded hydrogels in laboratory beagle dogsSurgical follow-up

IVD integrityExtracellular matrix metabolismCOX pathway and PGE2 levels

DiscussionIn vitro controlled release of CXB results into sustained suppression of PGE2 levels associated with inflammationSafe subcutaneous and intradiscal application of the thermoreversible pNIPAAM MgFe-LDH hydrogelSafe intradiscal application of the thermoreversible pNIPAAM MgFe-LDH hydrogel loaded with a wide range of CXB dosagesControlled release of CXB in mildly degenerated IVDs had a limited effect on PGE2 levels

ConclusionsAdditional filesAbbreviationsCompeting interestsAuthors’ contributionsAcknowledgementsAuthor detailsReferences

Documents

Biocompatibility and intradiscal application of a ...Introduction Chronic low back pain is a debilitating disorder associ-ated with intervertebral disc (IVD) degeneration [1]. As the